Multiple Zwave commands per Action causing lag

For a while now, I have been experiencing periods of high latency on events in response to button presses. In the logs, I see multiple occurrences of the same messages when this is happening:


In this example, device 65 is a GE 14295 Z+ dimmer. App 1258 is a simple button controller:
and dev 1686 is a motion detector connected to a Zooz ZEN17. Note that the lag between the last double tap message and the dim commend is 22 seconds. Any thoughts on what is causing this? I'm running a C-5 on 2.3.0.121, I experience these types of lags on other button commands as well, but not consistently.

How's your zwave mesh look? Even beyond that it's kind of strange behavior - especially since the office light switch shows physical events as well as digital. What is app 1258? And if you go look at device 65 what apps does it say the devcie is in use by?

Mesh looks okay insofar as I can see via PC Controller, since I have a C-5. No ghosts, tons of repeaters (mostly GE and Zooz wall switches), hub is centrally located in a 2,000 foot ranch. The only thing I can think of which might contribute is that 5 feet from the hub, I have a Honeywell T6 on a C-wire (acting as a repeater). The topology diagram in PC Controller shows that the thermostat has a huge number of neighbors, even more than the hub does. I am considering excluding and reincluding the thermostat on battery to make it nonrepeating then seeing how things settle out. Device 65 looks like this:

I'm puzzled as to what is causing the office light to be turned on and off physically... mostly that just means "external to HE" rather than literally a finger press. Could Google Home somehow be sending spurious on/off commands? Maybe a routine?

I did actually physically double tap the switch up one time. The associated button controller associated that with setting the office light to Dim 15%. The questions is why it logged so many times, and why there was a 20 second lag until the switch actually dimmed and then sent back multiple confirmations.

Is it possible that device 65 has simply gone bad? Maybe an exclusion, factory reset, and re-inclusion? Though you did say this was just an example.